ContextIsolation: Limits on Legacy Node Integration in Electron v15+

Goal

Developers migrating an Electron application to v15 or later face a choice: keep contextIsolation disabled to preserve legacy code that relies on nodeIntegration, or enable it for the security benefits it offers. The core question is how this decision affects existing renderer scripts that directly access Node.js APIs.

Uncertainty

Electron’s documentation states that contextIsolation defaults to true in v15+, but the impact on legacy code is not fully documented. Enabling isolation requires a preload script that uses contextBridge.exposeInMainWorld, which can break code that previously called require('fs') or other Node modules directly. There is no clear guidance on how to assess whether a renderer script can operate under isolation without modification.

Specific Questions

• How can a developer programmatically detect whether a particular renderer script depends on Node.js APIs that would be blocked when contextIsolation is true?
• What are the recommended patterns for exposing only the necessary Node functionality via contextBridge while keeping the rest of the renderer isolated?
• Is it possible to enable contextIsolation selectively per BrowserWindow instance so that legacy windows remain functional while new windows adopt isolation?